TJAAG - Research & Analysis

Holden Bruce

For the TJAAG Research & Data project, I will be focused on modeling the relationship between geography, wealth, and admissions.

Important Consideration

The data is inconsistent. Thomas Jefferson has only released middle school specific admissions data up until the Class of 2022 (2018 data) while reliable family income data from the Census is available up until 2019.

I do a few things in here to maintain some sort of consistency. Noteably: for admissions statistics, I reuse 2018 (Class of 2020) data when plotting against family income data.

import geopandas as gpd
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from mpl_toolkits.axes_grid1 import make_axes_locatable
import plotly.graph_objects as go
import plotly.express as px
from plotly.subplots import make_subplots

from IPython.display import HTML

middle_schools_areas = gpd.read_file('middle_schools_areas_jan31.shp')
middle_schools_areas

	SCHOOL_NAM	CITY	ZIP	Shape__Are	Shape__Len	YEAR	FAM_INCOME	PERC_POVER	TOT_ADMIT	AAP	geometry
0	CARSON	HERNDON	20171	4.090735e+08	140426.247323	2019	187130.689467	0.033291	78	YES	POLYGON ((-77.38339 38.95260, -77.38322 38.952...
1	COOPER	MCLEAN	22101	1.230428e+09	236742.140726	2019	351033.548622	0.023015	8	YES	POLYGON ((-77.32833 39.05769, -77.32816 39.057...
2	FRANKLIN	CHANTILLY	20151	4.057657e+08	154918.727298	2019	208201.580067	0.016814	0	NO	POLYGON ((-77.41445 38.92799, -77.41432 38.927...
3	FROST	FAIRFAX	22032	4.923730e+08	175702.028789	2019	192970.712256	0.016958	22	YES	POLYGON ((-77.23408 38.86523, -77.23427 38.864...
4	GLASGOW	ALEXANDRIA	22312	2.349001e+08	80961.006752	2019	132921.286973	0.103446	0	YES	POLYGON ((-77.16025 38.87394, -77.15826 38.873...
...	...	...	...	...	...	...	...	...	...	...	...
151	SOUTH COUNTY	LORTON	22079	8.878935e+08	223476.603738	2014	173697.751337	0.000000	5	YES	POLYGON ((-77.27914 38.74627, -77.27911 38.746...
152	STONE	CENTREVILLE	20120	5.508739e+08	117370.441464	2014	144352.026112	0.000000	4	NO	POLYGON ((-77.44312 38.84116, -77.44388 38.840...
153	THOREAU	VIENNA	22180	4.796591e+08	159064.168701	2014	180484.548503	0.000000	2	NO	POLYGON ((-77.29700 38.94074, -77.29742 38.940...
154	TWAIN	ALEXANDRIA	22310	2.417610e+08	84052.257861	2014	139677.655842	0.000000	11	YES	POLYGON ((-77.10140 38.80352, -77.10136 38.803...
155	WHITMAN	ALEXANDRIA	22306	6.397214e+08	181002.693552	2014	113918.763713	0.000000	1	NO	POLYGON ((-77.12013 38.76260, -77.11206 38.761...

156 rows × 11 columns

Total Students Admitted to TJ by Middle School

fig = px.bar(
    middle_schools_areas,
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'TOT_ADMIT'],
    title='Bar Plot of Admssions Statistics for 2017'
)
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
fig.show()

# from IPython.display import IFrame
# IFrame('https://arcg.is/1Cu85W0', width=800, height=500)

First step:

Plot the admissions data each year against the average family income for that middle school.

Important consideration:

I used block group census data for family income and overlayed the geometry of zones middle school attendance zones to aggregate the block group data up. I then divided by the number of families in that region to create an estimate of the average aggregate family income for a middle school student at each middle school.

While this is as good an estimate of middle school family wealth as we can get, it is still an imperfect estimate. Please keep in mind that because family-specific income data is not publicly available, take these findings with a grain of salt and understand that averages can always be misleading. A few wealthy families can skew the data.

A bit of hope:

The reason why I am not too concerned with the potential impact of using averages in this family income estimate is because the underlying assumption we are testing in this project is whether family income disproportionately impacts admissions data. If there are a few super wealthy families in a town and that is throwing off the average but those family's kids are getting admitted to TJ, then it doesn't really matter. Unfortunately, we cannot know which family's are gaining admission and what their family income is, so it is a moot point. An average of the entire region is the best we can do.

Room for improvement:

Using weighted averages of the various income brackets (i.e. \$0-\\$40k, \$40000-\\$70k, \$70k-\\$100k, etc) could better account for the true distribution of wealth in the middle school region.

Final reason for caring about averages of family income:

Public schools' funding is based in part on the income taxes paid by the local region. So, we expect a wealthy neighborhood to have a better funded school than a poor neighborhood.

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'TOT_ADMIT'], 
        marker=dict(
            size=15
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2019"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'TOT_ADMIT'], 
        marker=dict(
            size=10
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2019"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'TOT_ADMIT'], 
        marker=dict(
            size=10
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2018"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'TOT_ADMIT'], 
        marker=dict(
            size=10
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2017"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'TOT_ADMIT'], 
        marker=dict(
            size=10
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2016"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'TOT_ADMIT'], 
        marker=dict(
            size=10
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2015"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

# Create figure with secondary y-axis
fig = make_subplots(specs=[[{"secondary_y": True}]])
# Add traces
fig.add_trace(
    go.Scatter(
        mode='markers',
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'SCHOOL_NAM'], 
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'TOT_ADMIT'], 
        marker=dict(
            size=10
        ),
        name="Scatter Plot Admissions Data"),
    secondary_y=True,
)
fig.add_trace(
    go.Bar(
        x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'SCHOOL_NAM'],
        y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'FAM_INCOME'], 
        name="Bar Chart Income Data"),
    secondary_y=False,
)
# Add figure title
fig.update_layout(
    title_text="Plot of the Average Family Income and Admissions Rates for each Middle School in Fairfax County - 2014"
)
# Set x-axis title
fig.update_xaxes(title_text="Middle School Name")
# Set y-axes titles
fig.update_yaxes(title_text="<b>primary</b> Scatter Plot Admissions Data", secondary_y=True)
fig.update_yaxes(title_text="<b>secondary</b> Bar Chart Income Data",tickprefix="$", secondary_y=False)
fig.show()

Major Takeaways

While I assumed that there would be a direct corrletion between family income and admissions rates, that does not seem to be the case. There are several examples of wealth regions producing low admissions numbers.

From there, I explored some of the other factors that might contribute to a middle schooler's success. After reviewing the admissions data, it became clear to me that the level of academic rigor was prioritized in the admissions process. Middle School students who weren't taking advanced courses were by and large not even being considered in the application.

Why that matters

With that in mind, I looked at which school offer Advanced Academic Placement courses. The graph comparing admissions rates between the schools that did and did not offer AAP is stark.

middle_schools_areas
# now we can add the slider to the plot where each year is represented at diferent points on the slider
# https://stackoverflow.com/questions/59242519/plotly-how-to-use-multiple-datasets-with-plotly-express-bar-chart
# this is a simple example just using a bar graph
import plotly.express as px
# middle_schools_areas.TOT_ADMIT.max()
# middle_schools_areas.ZIP = middle_schools_areas.ZIP.astype('string') #changing type from num to zip so that color is discrete instead of continuous ... https://plotly.com/python/discrete-color/
fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2019'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2018','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2018'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2017','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2017'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2019','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2019'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2016','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2016'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2015','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2015'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2014'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

fig = px.bar(
    middle_schools_areas, 
    x=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'SCHOOL_NAM'], 
    y=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'TOT_ADMIT'],
    # color='AAP')
    animation_frame=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'YEAR'], 
    animation_group=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014', 'SCHOOL_NAM'], 
    range_y=[0,100],
    color=middle_schools_areas.loc[middle_schools_areas['YEAR']=='2014','AAP'],
    title='Bar Plot of the Admissions, separated by whether Middle School offers AAP - 2014'
    )
fig.update_layout(
    xaxis_title='Middle School Name',
    yaxis_title='Number of Students Admitted'
)
# fig.add_scatter(middle_schools_areas, x="SCHOOL_NAM", y="TOT_ADMIT")
fig.show()

Important considerations for the future:

Longfellow makes sense to me, Cooper does not. What is going on there? Exploring the programs available at the elementary schools that feed into those middle schools might illuminate some things that are currently hidden. AAP is clearly an important indicator, are there elementary school equivalents that impact this even more? Being able to extend the timeline back in time (earlier in the child's academic life) could be an interesting path to explore.

What percentage of the families living in the middle school attendance zones are sending their kids to public schools? The Cooper Middle School zone is one of the wealthiest but had a very low admissions rate for TJ, are the wealthy families in that region sending their kids to private school, leaving a poorer student body than what is represented in the data?

Did not account for Loudoun County, Prince William County, or any other students who entered TJ without going through the FCPS system.